基于支持向量机的保健品消费者行为研究

发布时间：2018-02-26 09:13

本文关键词： 消费者行为支持向量机潜在购买力保健品仿真技术　出处：《安徽理工大学》2013年硕士论文　论文类型：学位论文

【摘要】：近年来,保健品行业和消费者行为均为学术研究的热点。如何利用现有的技术和理论,实现数据仿真,构建行为模型,挖掘潜在保健品消费群体和潜在购买力将成为保健品行业的重要发展方向。本文中的保健品包括食用的蜂制品、维生素等一般营养品,具有保健功能的保健药品、药膳和保健化妆品。保健品的消费情况不同于一般的生活日用品,而且其消费更易受更多的因素影响,需要探索一条适合保健品消费者行为的研究途径。文章首先简单介绍了课题研究的背景、意义、目的和方法等,并论述了国内外消费者行为的研究现状,指出国内研究起步晚、研究成果不多、研究方法有待创新。分析了我国保健品市场经历了起步阶段、快速发展阶段、萧条发展阶段和复苏调整阶段。在短短十几年时间里,保健品已经迅速发展成为一个独特的、备受各界瞩目的热点产业,成为中国工业经济新的增长点和国民经济的新兴行业。尽管我国保健品行业发展潜力大、产品质量和科技含量不断提高、具备一定的规模化,但仍面临着一些挑战。保健品未来的发展将呈现消费者群体扩大化、营销模式专营化、产品材料多样化、产品类型多元化,市场监控规范化的趋势。文中研究了支持向量机和BP神经网络算法的基本原理和学习训练过程,分析了两种算法的优缺点。参考国内外研究成果和问卷调查的基本原则,设计了适合保健品消费者行为研究的问卷,并针对有购买保健品意向的网民消费者发布问卷,收集有效问卷。利用SPSS19.0对收回的问卷信息进行描述性分析和信度、效度分析,表明问卷具有一定的实用性。本文将处理后的影响保健品消费者行为的各种重要因素即性别、年龄、收入等,通过仿真环境,进行回归预测,并对具体流程做了详细的描述。实验表明,支持向量机算法在保健品消费者的购买能力和实际购买的预测中,比神经网络算法具有更高的准确性。最后得到以下结论：支持向量机算法在解决小样本、高维模式和非线性问题上具有较好的逼近能力和泛化能力；影响保健品消费者行为的主要因素有收入、节省程度、职位类别和婚姻状况等；有购买意向的保健品消费者平均的实际购买占购买能力的2/3,具有一定的可挖掘性,且购买能力越高的消费者,其可挖掘性越大。
[Abstract]:In recent years, health products industry and consumer behavior have been the hot topics of academic research. How to use existing technology and theory to realize data simulation and build behavior model, It will be an important development direction for the health products industry to tap the consumption groups and potential purchasing power of potential health products. The health products in this paper include bee products, vitamins and other general nutrients, health drugs with health care functions. The consumption of health products is different from that of general daily necessities, and its consumption is more easily affected by more factors. Therefore, it is necessary to explore a research approach suitable for the consumer behavior of health products. Firstly, the paper briefly introduces the background, significance, purpose and methods of the research, and discusses the current research situation of consumer behavior at home and abroad, pointing out that the domestic research started late, and the research results are few. The research methods need to be innovated. It is analyzed that the health care products market in our country has experienced the initial stage, the rapid development stage, the depression development stage and the recovery adjustment stage. In a short period of ten years, the health care products have developed rapidly into a unique, Hot industries, which have attracted much attention from all walks of life, have become a new growth point of China's industrial economy and a new industry of the national economy. Despite the great potential for development of the health products industry in China, the quality of products and the scientific and technological content have been continuously improved, with a certain scale. However, there are still some challenges. The future development of health products will present the expansion of consumer groups, the monopoly of marketing mode, the diversification of product materials, and the diversification of product types. This paper studies the basic principle and learning and training process of support vector machine and BP neural network algorithm, analyzes the advantages and disadvantages of the two algorithms. This paper designed a questionnaire suitable for the study of consumer behavior of health products, and issued a questionnaire for consumers with intention to purchase health care products, and collected valid questionnaires. SPSS19.0 was used to analyze the information collected from the questionnaires in terms of descriptive analysis, reliability and validity analysis. The results show that the questionnaire has a certain practicability. In this paper, all kinds of important factors, such as gender, age, income and so on, which affect the consumer behavior of health products after treatment, are predicted by the simulation environment. The experimental results show that the SVM algorithm is more accurate than the neural network algorithm in predicting the purchasing ability and actual purchase of health products consumers. Finally, the following conclusions are obtained: the SVM algorithm has better approximation and generalization ability in solving small samples, high dimensional patterns and nonlinear problems, and the main factors affecting the consumer behavior of health products are income, saving degree, etc. The average actual purchase of health care products with purchase intention accounts for 2 / 3 of the purchasing capacity, which has a certain degree of diggability, and the higher the purchasing power, the greater the excavability of consumers.
【学位授予单位】：安徽理工大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP181;F719

【参考文献】